CN110943745A

CN110943745A - Polarization code BP decoding method and system for early stopping iteration output result

Info

Publication number: CN110943745A
Application number: CN201911204977.4A
Authority: CN
Inventors: 贾光帅; 许聪; 李泉泉; 顾大晔; 叶文静; 陶子然; 尹健; 李岩
Original assignee: Anhui Core Century Technology Co ltd; CETC 38 Research Institute
Current assignee: Anhui Core Century Technology Co ltd; CETC 38 Research Institute
Priority date: 2019-11-29
Filing date: 2019-11-29
Publication date: 2020-03-31
Anticipated expiration: 2039-11-29
Also published as: CN110943745B

Abstract

The invention provides a method and a system for decoding a polarization code BP (Back propagation) of a result of early termination iteration output, which comprise the following steps: s1, screening channels; s2, constructing an information block and carrying out coding transmission: the information block U comprises K-K_CRCAn information bit; k is the information length of each polar code encoding module, K_CRCThe number of the added CRC check bits; CRC check bits are added to the information bits before encoding, then K-K_CRCAdding one information bit to K_CRCThe CRC check bit enters a polar code encoder; s3, receiving the code word signal transmitted by the information block through coding and decoding: in the aspect of decoding, BP decoding is carried out on the received N code words; the decoding estimation value of the information block U can be obtained every time BP decoding iteration is carried out

For the obtained decoding estimated value

Carrying out G_NMatrix check and CRC check: if the estimated information bits simultaneously satisfy G_NIf the matrix check and the CRC check are carried out, the decoding can be stopped; otherwise, the next round of BP iteration will continue until G is satisfied at the same time_NAnd carrying out matrix check and CRC check or reaching the set maximum iteration number.

Description

Polarization code BP decoding method and system for early stopping iteration output result

Technical Field

The invention relates to the technical field of fading channel communication, in particular to a polarization code BP decoding method and a system for terminating iteration output results in advance.

Background

The polar code proposed by Arikan has been demonstrated to be a coding scheme that achieves channel capacity in arbitrary binary input discrete memoryless channels (B-DMCs), which is decoded using the Serial Cancellation (SC) algorithm, and has low coding and decoding complexity. SC algorithm decoding is a special example of Belief Propagation (BP) algorithm decoding. In general, BP decoding performance is better than SC decoding in additive white gaussian noise channel (AWGN) polar codes.

However, the existing early termination criterion (referred to as the G _ N matrix termination criterion) is not suitable for the environment of BP decoding in a fading channel, and the phenomenon of outputting wrong decoding due to early termination iteration occurs. Under the same conditions, the polar code BP decoding performance in a Fading channel (Fading channel) no longer has an advantage over the SC decoding performance. In addition, the conventional early termination method may have errors in a fading channel due to early termination of the iteration when the termination condition is satisfied. Therefore, how to provide a polar code BP decoding method for terminating an iterative output result in advance without increasing decoding complexity is a technical problem which needs to be solved urgently.

Disclosure of Invention

The invention aims to solve the defect that the belief propagation algorithm decoding performance of the polar code under a fading channel is not ideal in the prior art, and provides a polar code BP decoding method for terminating an iteration output result in advance to solve the problem. .

The invention solves the technical problems through the following technical means:

a polarization code BP decoding method for stopping iteration output results in advance comprises the following steps:

s1, screening channels: calculating a channel parameter matrix P_eAccording to a channel parameter matrix P_eScreening out an information bit channel and a frozen bit channel, putting the information bit channel into a set I, and putting the frozen bit channel into a set F;

s2, constructing an information block and carrying out coding transmission: the information block U comprises K-K_CRCAn information bit; k is the information length of each polar code encoding module, K_CRCThe number of the added CRC check bits; CRC check bits are added to the information bits before encoding, then K-K_CRCAdding one information bit to K_CRCThe CRC check bit enters a polar code encoder;

s3, receiving the code word signal transmitted by the information block through coding and decoding: in the aspect of decoding, BP decoding is carried out on the received N code words; the decoding estimation value of the information block U can be obtained every time BP decoding iteration is carried out

For the obtained decoding estimated value

Carrying out G_NMatrix check and CRC check: if the estimated information bits simultaneously satisfy G_NIf the matrix check and the CRC check are carried out, the decoding can be stopped; otherwise, it will continueThe next round of BP iteration until G is satisfied simultaneously_NAnd carrying out matrix check and CRC check or reaching the set maximum iteration number.

Preferably, the step S1 specifically includes the following steps:

s11, aiming at all bit channels, obtaining a channel parameter matrix P containing error probability values of all split channels by using a construction method_e；

S12, according to the channel parameter matrix P_eSelecting an information bit channel set I and a frozen bit channel set F, which specifically comprises the following steps:

s121, comparing channel parameter matrix P of bit channel_eThe elements in the channel parameter matrix are sorted in ascending order according to the error probability of each split channel to obtain a sorted channel parameter matrix P_tempAnd find P_tempEach element in P_eTo obtain an index matrix P_idx；

S122, extracting the index matrix P_idxThe first K elements in the code are sorted in ascending order and put into an information bit channel set I, wherein K is the information bit number of the polarization code; will P_idxThe remaining elements are sorted in ascending order and placed into a frozen bit channel set F.

Preferably, the step S2 specifically includes the following steps:

s21, dividing the message bits input in sequence into a plurality of information frames, wherein each information frame comprises K information bits;

s22, in each information frame, the first K-K is_CRCPutting information bits into an information block U;

s23, adding CRC check codes and transmitting: and adding a CRC check code at the tail end of the information bit according to the CRC polynomial to the information block U by adopting a CRC generator polynomial of the international standard for coding transmission.

Preferably, the step S3 specifically includes the following steps:

s31, for each information frame, the decoder receives the code word and the frozen bit set of the information block U

Generating a decoded estimate of U

S32、G_NThe matrix check module receives the estimated value of decoding

And returning a check result, and analyzing by the decoder according to the check result;

s33, decoding the estimated value

G of (A)_NIf the matrix check result is correct, the decoder continues to decode the estimated value

And (3) performing CRC check on the CRC module, and performing corresponding processing according to the check result, wherein the specific steps are as follows:

s331, receiving the decoding estimated value by the CRC check module

Returning a check result, and selecting the next operation again by the decoder according to the check result;

s332, if the CRC check result is correct, the decoder terminates iteration in advance, the decoding result of the information block is directly output, and the information frame decoding is finished;

s333, if the CRC result is wrong, the iterative decoding fails, and the information frame is iterated next time until G is satisfied_NMatrix checking and CRC checking or reaching a set maximum iteration number;

s334, if the estimated value is decoded

G of (A)_NIf the matrix check result is wrong, the iterative decoding fails, and the information frame is iterated next time until G is satisfied simultaneously_NAnd carrying out matrix check and CRC check or reaching the set maximum iteration number.

Preferably, the set maximum number of iterations is generally set to 70.

Preferably, the decoder is a BP decoder.

Based on the method, the invention also provides a polarization code BP decoding system for stopping iteration output results in advance, which comprises

A screening channel module: calculating a channel parameter matrix P_eAccording to a channel parameter matrix P_eScreening out an information bit channel and a frozen bit channel, putting the information bit channel into a set I, and putting the frozen bit channel into a set F;

the information quick construction module: the method is used for constructing information blocks and carrying out coding transmission, and specifically comprises the following steps: the information block U comprises K-K_CRCAn information bit; k is the information length of each polar code encoding module, K_CRCThe number of the added CRC check bits; CRC check bits are added to the information bits before encoding, then K-K_CRCAdding one information bit to K_CRCThe CRC check bit enters a polar code encoder;

a decoding module: the method is used for receiving and decoding a code word signal transmitted by an information block through coding, and specifically comprises the following steps: BP decoding is carried out on the received N code words; the decoding estimation value of the information block U can be obtained every time BP decoding iteration is carried out

For the obtained decoding estimated value

Preferably, the step of screening channels by the channel screening module specifically includes the following steps:

s11, obtaining the channel containing each split signal by using a construction method aiming at all bit channelsChannel parameter matrix P of channel error probability values_e；

Preferably, the information fast constructing module specifically executes the following process:

Preferably, the decoding module specifically executes the following process:

Generating a decoded estimate of U

S32、G_NThe matrix check module receives the estimated value of decoding

s33, decoding the estimated value

s331, receiving the decoding estimated value by the CRC check module

s334, if the estimated value is decoded

The invention has the advantages that:

the invention relates to a polarization code belief propagation decoding algorithm method suitable for a termination iteration criterion in advance in a fading channel. The error code performance of the polar code BP decoding is improved, the decoding complexity is not increased, and the practicability of a polar code decoding algorithm is greatly improved.

The invention satisfies G simultaneously in the decoding iteration process_NThe two check conditions of the matrix check and the CRC check are used for terminating iteration in advance so as to improve the decoding performance. In the encoding stage, each information block adds CRC check bits. In the decoding stage, estimated information bits occur

Satisfies G_NAnd when the matrix is checked and CRC check is met, ending the decoding iteration process in advance and outputting the decoding. Errors due to premature termination do not occur in the proposed termination rule except for uncorrectable information blocks.

Unlike the conventional early termination method, errors due to early termination do not occur in the proposed termination method except for uncorrectable information blocks. May be abbreviated as G_Nmatrix-CRC termination criteria method.

The simulation results of the present invention show that G is used_NIn the matrix-CRC termination criterion method (the early termination iteration method described in the present invention), in a fading channel, the code length N is 256, the code rate R is 0.36, and the maximum number of iterations is 70, the error code performance of BP decoding is already better than SC decoding. In addition, it is noted that G is used_NBP decoding of the matrix-CRC termination criterion method improves decoding efficiency at the same time.

Drawings

FIG. 1 is a block flow diagram of a method in an embodiment of the invention;

FIG. 2 shows a diagram of G according to an embodiment of the present invention_NCoding a decoding system model diagram by a matrix-CRC termination criterion method;

FIG. 3 is a graph comparing the error performance of BP decoding and SC decoding with a maximum number of iterations of 30 in AWGN according to an embodiment of the present invention;

fig. 4 is a diagram illustrating comparison of the error code performance of BP decoding and SC decoding with a maximum number of iterations of 30 in a fading channel according to an embodiment of the present invention;

FIG. 5 shows an embodiment of the present invention employing G in a fading channel_Nmatrix-CRC termination criteriaThe error code performance comparison graph of BP decoding and other decoding modes is obtained;

FIG. 6 shows an embodiment of the present invention employing G in a fading channel_NThe average iteration number graph when the matrix-CRC termination criterion method BP decodes;

FIG. 7 shows an embodiment of the present invention employing G in a fading channel_NAn iteration condition analysis graph in the matrix termination criterion method BP decoding process;

FIG. 8 shows an embodiment of the present invention employing G in a fading channel_NAnd (3) an iteration condition analysis diagram in BP decoding of the matrix-CRC termination criterion method.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The invention first calculates the channel parameter P by the existing construction method, such as the Tal-Vardy algorithm or the Gaussian approximation method_eThen, selecting information bits and freezing bits according to the channel parameters; and then adding CRC check codes to the information frames. Based on BP decoder and G_NMatrix and CRC checks (CRC-16 is used here, the polynomial of which is poly [10001000000100001 ]]) And respectively implementing different decoding strategies on the results, and finally outputting the decoding results.

As shown in fig. 1, the method for decoding a polarization code BP that terminates an iterative output result in advance according to the present invention includes the following steps:

in the first step, channels are screened.

The channel parameter P is calculated by means of known construction methods, such as the Tal-Vardy algorithm or the Gaussian approximation method_eAccording to a channel parameter matrix P_eScreening out information bit channels and freezing bit channels, putting the information bit channels into a set I, and freezing the bit channelsThe tracks are placed in set F. The method comprises the following specific steps:

(1) aiming at all bit channels, a channel parameter matrix P containing error probability values of all split channels is obtained by utilizing the existing Tal-Vardy algorithm or Gaussian approximation method_e。

(2) According to the channel parameter matrix P_eSelecting an information bit channel set I and a frozen bit channel set F, which specifically comprises the following steps:

A1) channel parameter matrix P for bit channel_eCalling the sort function in matlab according to the existing mode to store P_eThe elements in the list are sorted in ascending order according to the error probability of each split channel, and the function returns the sorted channel parameter matrix P_tempAnd P_tempEach element in P_ePosition index matrix P in (1)_idx。

A2) Fetching index matrix P_idxThe first K elements (K is the number of information bits of the polar code, i.e. the number of information bits in a polar code encoded information block, K_CRCFor the number of CRC check bits in each block) into a set I of information bit channels in ascending order, P_idxThe remaining elements are sorted in ascending order and placed into a frozen bit channel set F.

In the second step, the information block is constructed and transmitted as shown in fig. 2.

Firstly, message bits input sequentially are divided into a plurality of information frames, and each information frame comprises K information bits.

Secondly, each information block U contains K-K_CRCAn information bit, K being the length of each information block (number of information bits of the polar code), K_CRCIs the number of CRC bits added.

And finally, adding a CRC (cyclic redundancy check) code to each information block, and carrying out coding transmission.

The specific steps of constructing the information block and performing coding transmission are as follows:

B1) the method comprises the steps of dividing sequentially input message bits into a plurality of information frames, wherein each information frame comprises K information bits.

B2) For each information frame, it is preceded by K-K_CRCInformation ratioBits are placed in the information block U.

B3) And adding a CRC check code and transmitting. And (3) adopting a CRC generator polynomial of the international standard, and for the information block U, respectively adding CRC check codes at the tail ends of the information bits according to the CRC polynomial and then transmitting the information block U.

And thirdly, receiving the code word signal of the information block which is transmitted by coding and carrying out a decoding process. Receiving and decoding a code word signal transmitted by the information block through coding; decoding the estimated value

To G_NMatrix check and CRC check module, pass G_NReturning a check result by the matrix check and the CRC check; according to G_NAnd selecting whether to perform a new decoding or not according to the check results returned by the matrix check and the CRC check.

The coded codeword signal is transmitted in a symmetric B-DMC channel W and observed at the receiving end. The receiving end collects the code words with the length of N each time. The decoder generates a decoded estimate for each information block U

Then sent to G_NMatrix check and CRC check, G possible_NThe matrix check and CRC check results are:

case 1: g_NThe matrix check is correct and the CRC check is correct.

Case 2: g_NThe matrix check is correct and the CRC check is incorrect.

Case 3: g_NThe matrix checks for errors.

For case 1, due to G_NIf the matrix check is correct and the CRC check is correct, judging that the decoding result is correct, terminating iteration in advance and outputting a decoding; for case 2, G_NThe matrix check is correct, but the CRC check is wrong, so the next round of BP iteration needs to be executed until G is satisfied at the same time_NMatrix checks and CRC checks or iterations to a maximum number of iterations. For case 3, due to G_NIf the matrix check is wrong, CRC check is not needed, and the decoder directly performs the next round of BP iteration until the next round of BP iteration is performed simultaneouslySatisfies G_NMatrix checks and CRC checks or iterations to a maximum number of iterations.

When a new round of coding occurs, its more detailed coding process is described as follows:

the specific steps of receiving the code word signal transmitted by the information block after coding and carrying out the decoding process are as follows:

(1) for each information frame, the decoder receives the code word of the coded transmission of the information block U and the frozen bit set

Generating a decoded estimate of U

(2)G_NMatrix check and CRC check module reception

And returning a check result, and analyzing by the decoder according to the check result.

(3) If the estimated value is decoded

G of (A)_NIf the matrix check result is correct, then performing CRC check, and performing corresponding processing according to the check result, specifically comprising the following steps:

C1) if the estimated value is decoded

If the CRC check result in (1) is correct, the decoder directly outputs the decoding result of the information block, and the decoding of the information frame is finished, which corresponds to the case 1 described above.

C2) If the estimated value is decoded

If the CRC result is wrong, the decoder directly executes a new round of BP decoding until G is satisfied simultaneously_NThe matrix check and CRC check or iteration to the maximum number of iterations corresponds to case 2 above.

(4) If the estimated value is decoded

G of (A)_NIf the matrix check result is wrong, the decoder directly executes a new round of BP decoding until G is satisfied simultaneously_NThe matrix check and CRC check or iteration to the maximum number of iterations corresponds to case 3 above.

G described for the invention_Nmatrix-CRC termination method, G_NBoth the matrix check and the CRC check are used to detect whether the codeword is valid, which is equivalent to determining whether the estimated information bits are correct. This double valid criterion results in that no errors due to unreliable early termination occur except for uncorrectable information blocks.

As shown in fig. 3, fig. 3 is a graph comparing the error performance of BP decoding and SC decoding in AWGN, where the maximum number of iterations originally set to be 30, and K represents the number of information bits in one polar code encoded information block (K, K is N × R, and N is 256, R is 0.36, and K is 92, calculated according to the difference between the code length and the code rate), it can be found that the original BP decoding performance in AWGN is superior to SC decoding.

As shown in fig. 4, fig. 4 is a graph comparing the bit error performance of BP decoding and SC decoding originally set with the maximum iteration number of 30 in the fading channel, where K represents the number of information bits in one polar code encoded information block (K, K — N × R is calculated according to the difference between the code length and the code rate, and N — 256, R — 0.36, and K — 92 in the simulation of the present invention), it can be found that the original BP decoding performance in the fading channel no longer has better performance than SC decoding.

FIG. 5 shows the use of G in a fading channel according to the present invention_NThe error code performance comparison chart of the matrix-CRC termination criterion method BP decoding and other decoding modes. The simulation parameters are as follows: the number of information bits in a block of polar coded information is denoted by K (K, K-N R, calculated from the difference between code length and code rate; N-256 in the simulation of the invention, R-0.36, K-92), K_CRCIndicating the CRC check bits in each blockThe CRC check bits are part of K information bits. Thus, each polar code information block has a K-K_CRCNew information bits. And adding a CRC check code, then coding and transmitting in a fading channel. The decoding process is already given in the third step above. The K92 information bits contain a 16-bit CRC check code. As can be seen from FIG. 5, G is used_NBP decoding performance of the matrix-CRC termination criterion method is obviously superior to SC and G_NThe BP decoding of the matrix termination criterion method is superior to BP decoding which only improves the decoding performance by increasing the maximum iteration number to 70.

FIG. 6 shows the use of G according to the invention_NComparing the average iteration times of the matrix-CRC termination criterion method with that of the CRC; the simulation parameters of fig. 6 are identical to those of fig. 5. As can be seen from FIG. 6, G_NThe average iteration times of the matrix-CRC termination criterion method are greatly reduced, and the decoding efficiency is improved.

FIG. 7 shows the use of G in a fading channel according to the present invention_NAn iteration condition analysis graph in the matrix termination criterion method BP decoding process; the simulation parameters of fig. 7 are consistent with fig. 5. In the figure, black circles indicate each group of information blocks, and red crosses indicate information blocks with decoding errors. As can be seen from the figure, G is used_NThe matrix termination criterion method may occur in part because of early termination of the iteration, outputting erroneous decoding results.

FIG. 8 shows the use of G according to the invention_NA matrix-CRC termination criterion method iteration condition analysis graph; the simulation parameters of fig. 8 are consistent with fig. 5. As can be seen from the figure, G is used_NThe matrix-CRC termination criteria method has no error cases due to early termination of iterations, except for uncorrectable error blocks.

The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. A polarization code BP decoding method for stopping iteration output results in advance is characterized by comprising the following steps:

For the obtained decoding estimated value

2. The method as claimed in claim 1, wherein the step S1 specifically includes the following steps:

s11, aiming at allBit channel, obtaining channel parameter matrix P containing error probability value of each split channel by using construction method_e；

3. The method as claimed in claim 1, wherein the step S2 specifically includes the following steps:

4. The method as claimed in claim 1, wherein the step S3 specifically includes the following steps:

Generating a decoded estimate of U

S32、G_NThe matrix check module receives the estimated value of decoding

s33, decoding the estimated value

s331, receiving the decoding estimated value by the CRC check module

s334, if the estimated value is decoded

G of (A)_NIf the matrix check result is wrong, the iterative decoding fails, and the information frame is processed for the next timeIterate until G is satisfied at the same time_NAnd carrying out matrix check and CRC check or reaching the set maximum iteration number.

5. The BP decoding method according to claim 4, wherein the maximum number of iterations is set to 70.

6. The BP decoding method according to claim 4, wherein the decoder is a BP decoder.

7. A polarization code BP decoding system for stopping iteration output results in advance is characterized in that: comprises that

For the obtained decoding estimated value

Carrying out G_NMatrix check and CRC check:if the estimated information bits simultaneously satisfy G_NIf the matrix check and the CRC check are carried out, the decoding can be stopped; otherwise, the next round of BP iteration will continue until G is satisfied at the same time_NAnd carrying out matrix check and CRC check or reaching the set maximum iteration number.

8. The BP decoding system according to claim 7, wherein the BP decoding system comprises: the channel screening module for screening channels specifically comprises the following steps:

9. The BP decoding system according to claim 7, wherein the information block constructing module specifically executes the following process:

10. The method of claim 7, wherein the decoding module performs the following steps:

Generating a decoded estimate of U

S32、G_NThe matrix check module receives the estimated value of decoding

s33, decoding the estimated value

s331, receiving the decoding estimated value by the CRC check module

s334, if the estimated value is decoded